This document discusses optimistic concurrency control, which is a concurrency control method that assumes transactions can frequently complete without interfering with each other. It operates by allowing transactions to access data without locking and validating for conflicts before committing. The validation checks if other transactions have read or written the same data. If a conflict is found, the transaction rolls back and restarts. The document outlines the basic algorithm, phases of transactions (read, validation, write), and advantages like low read wait time and easy recovery from deadlocks and disadvantages like potential for starvation and wasted resources if long transactions abort.

What's hot

Similar to Optimistic concurrency control in Distributed Systems

Recently uploaded

Optimistic concurrency control in Distributed Systems

• 1.
  Optimistic ConcurrencyControl- BYMRIDUL K. MISHRA(170303201015)
• 2.
  What is It?Concurrency control is a concept that is used to address conflicts with thesimultaneous accessing or altering of data that can occur with a multi-usersystem. Optimistic concurrency control (OCC) is a concurrency control method appliedto transactional systems. OCC assumes that multiple transactions can frequentlycomplete without interfering with each other. While running, transactions use dataresources without acquiring locks on those resources. Before committing, eachtransaction verifies that no other transaction has R/W the data it has R/W. If thecheck reveals conflicting modifications, the committing transaction rolls back andcan be restarted.
• 3.
   In contextof DS, it can be advantageous to apply to systems which are very largeand mostly involve resource accesses which don’t alter data in resource. In otherwords, this should be applied to systems that are immune to conflictingoperations. Its applied to systems that are willing to accept/tolerate temporary inconsistenciesand deadlocks.
• 4.
  Why is itused? It is expensive and complicated to implement locks on resources, and it alsoreduces the concurrency of operations. It neutralises the effects of conflicts(if any).
• 5.
  How is itimplemented? The basic algorithm for OCC was designed by H.T. Kung and John Robinson. Approach is called ”Optimistic”, because it is based on the observation that, inmost applications, the likelihood of two transactions accessing the same object islow. The approach “hopes” that conflicts do not occur and transactions are allowed toproceed as though there were no possibility of conflict. Objective is to minimize the time over which a given resource would beunavailable for use by other transactions
• 6.
  Algorithm Since readinga value or a pointer from a node can never cause a loss of integrity,reads are completely unrestricted Writes are highly restricted. Transactions consist of three phases: Read Phase: Resources can be read freely. All writes take place on local copiesof the object to be modified i.e. a local updated copy is created. Validations Phase: The step in which it is determined that the transaction willnot cause a loss of integrity. i.e. if any R-W or W-R anomaly occurs Write Phase: Copies are made global, if the data is consistent the file is updated.
• 7.
  Read & WritePhase Read is also referred to as the “Working Phase” Each transaction has a tentative version of each of the object that it updates READ operations are performed immediately WRITE operations record the new values of the objects as tentative values Two records are kept of the objects accessed within a transaction: a read set and awrite set
• 8.
   If validationsucceeds, then the transaction enters the write phase After write phase, all written values become “global” When a transaction completes, it will request its validation and write phases viatransactionEnd’ call
• 9.
  Validation Phase Usesa particularly strong form of validation, is also responsible for deadlockprevention. This is especially important with long-running transactions Method uses an overqualified update scheme to test whether the underlying datasource has been updated by another transaction since the beginning of thecurrent transaction Kung and Robinson employ Serial Equivalence for verifying the correctness ofconcurrent execution of transactions.
• 11.
  Validation Rules:T1 T2RuleWrite Read T2 must not access data being altered by T1.Read Write T1 must not read data being altered by T2.Write Write T2 must not alter data altered by T1 vice-versa.
• 12.
  Advantages The waitingtime to Read a resource is very low. Deadlocks can be very easily recovered(by validation phase). Read-only transactions can run concurrently with updating transactions withoutloss of database consistency. Never lead to cascaded aborts.
• 13.
  Disadvantages: Can leadto starvation of process that want to alter the contents of a resource. Aborting the long transactions in validation phase wastes a lot of systemresources. Since one transaction cant access resource of other ones, it limits concurrencywhen write function is involved.
• 14.
  Refernces Google.com Wikipedia
• 15.
  Thank You